Llama 3.2 vision 11B model for OCR task

#1
by Pavankumar03 - opened

Hello,
Can we use the Llama-3.2-11B-Vision-Instruct model for OCR task...??
If yes, help me with the quantization of this model for OCR task....

Unsloth AI org

Hello,
Can we use the Llama-3.2-11B-Vision-Instruct model for OCR task...??
If yes, help me with the quantization of this model for OCR task....

Absolutely there's notebook examples for vision models : https://unsloth.ai/blog/vision

Using this 4bit version will be perfect for it

yeah thankyou got it,
where can i get the notebook for quantizing that model(Llama 3.2 Vision Model)
I am just using that model of yours.
Can i get that code of quantizing process?

I'm using it to OCR vehicle VIN Codes. It's satisfactory, with just a few thousand training photos it's already 70% more effective than the base model.

Unsloth AI org

I'm using it to OCR vehicle VIN Codes. It's satisfactory, with just a few thousand training photos it's already 70% more effective than the base model.

Incredible glad to see it :)

Sign up or log in to comment